Introduction

In the Global North migrants’ working lives are routinely defined by the precarity of employment at the bottom of the labour market (Lewis et al., 2015). As a result, they usually face indecent working and living conditions and major vulnerability to economic recessions (ILO, 2012). In Spain, Moroccan and Ecuadorian immigrants are among the most disadvantaged ethnic groups in the occupational structure, with the lowest wages and the worst contractual conditions (Miguélez et al., 2011). These disadvantages, moreover, intersect with gender inequalities, restricting immigrant women’s options in a gendered labour market that reinforces their entrapment in its lowest positions (Parella, 2003). How are these phenomena explained?

Two main approaches have been used to explain immigrants’ disadvantage and ethnic inequalities in the labour market. First, human capital theory treats work as just another tradeable market commodity adjusted through the equilibrium between the supply and demand for labour. In the market, employees’ investments in education and training are rewarded by employers because of the higher productivity that they offer (Degenne and Forsé, 1999). From this perspective, inequalities in the labour market are explained mostly by differences in the individual’s decision to invest in education and training (Becker, 1983). In the case of migrants, Chiswick (1978) also considers the international transferability of human capital skills, namely language, job-related skills, labour market information, and credentials. These factors are expected to increase as time passes in the destination, therefore drawing a U-shaped pattern of occupational change over time in the host country (Chiswick et al., 2003).

Secondly, labour market stratification theory rejects the assumption that labour market divisions can be attributed mainly to inadequate levels of human capital or differences in productivity. Following some of the long-standing critiques of human capital theory (see Bowles and Gintis, 1975), labour market stratification theory claims that the concept perpetuates an overly simplistic view of the labour market based on individual choices and market rewards that end up blaming the victim, that is, those who are unsuccessful in the labour market. This approach, by contrast, highlights the role of the structure of labour demand instead. It considers that inequalities are fostered mainly through both formally and informally institutionalized policies and practices in labour markets and workplaces (Grimshaw et al., 2017). Labour market stratification theory also believes that, as a result, the labour market is divided into different segments with different structures and characteristics. On the one hand, a primary segment with high levels of security and stability, possibilities for promotion and prestigious jobs. On the other hand, a secondary segment with low wages, job instability, bad working conditions and a lack of career prospects (Piore, 1975; Grimshaw et al., 2017).

According to the labour market stratification theory, the labour market is stratified along ethnic lines in a way that pushes immigrants to occupy positions in the secondary segment, specifically in employment nichesFootnote 1 (Pajares, 2005; Miguélez et al., 2011). This theory acknowledges that transnational migrant processes and processes of exclusion practised by states often close down real and acceptable alternatives for migrants to engage in precarious and exploitative labour (Lewis et al., 2015). In Spain, in particular, the growth of immigration has both enabled and been promoted by the economic expansion of some labour-intensive industrial sectors. This process has been reinforced by Spanish migration policies by facilitating the placement of immigrants in ‘hard-to-fill jobs’ (that local workers are less inclined to do), and more broadly by creating a ‘discriminatory institutional framework’ that restricts access to residence and work permits, and places barriers to the recognition of academic titles (Cachón, 2009; Veira et al., 2011), particularly to immigrants from the Global South. It all contributes to placing such labour-intensive industrial sectors in the secondary labour market, characterized by low wages, limited contractual security, and little stability or possibilities for promotion (Martinez Veiga, 1997; Miguélez et al., 2011). These sectors have thus become ethnic occupational niches in which the jobs on offer are arduous, dangerous and precarious (Cachón, 2003, p. 38), ones rejected by the autochthonous population (Parella, 2003) and often situated in the informal economy (Pajares, 2005). Diverse mechanisms of ethnic and/or racial discrimination by employers and society as a whole contribute to binding immigrants within these sectors, with some occupations coming to be regarded as ‘immigrants’ jobs’ (Cachón, 2003; Parella, 2012). In Spain these ethnic occupational niches include construction, agriculture, hotels and catering, retail and domestic work (Pajares, 2010).

Finally, the labour market stratification theory also highlights the fact that concentration in a few employment sectors is even stronger for immigrant working women. In the context of a sexually segregated occupational structure, immigrant women also face gendered discrimination in their access to paid work that relegates them to the employment niches that Spanish women reject (Parella, 2003, p. 138). The growing influx of female migration has been closely linked to the expansion of the service sector, the mercantilisation of care (Parella, 2003) and the subsequent growth of employment opportunities in hotels, catering and particularly domestic work (Moreno-Colom and López-Roldán, 2018). As a result, immigrant women’s employment is strongly bounded by or concentrated in stigmatized employment niches, in positions that are unqualified, undervalued, and have markedly servile connotations (Moreno-Colom and López-Roldán, 2018).

There is also a third fertile stream of literature, which emphasizes the importance of seeking jobs through social networks in enabling individuals to access certain positions in the labour market (Granovetter, 1973, 1974; Lin, 1999). The social network perspective on the labour-market integration of migrants does not fit clearly into either of the two previous approaches. Indeed, this perspective makes a clear criticism of the human capital approach, as it recognizes that information on job opportunities does not flow properly on the labour market, stressing that mechanisms other than investment in education and training explain how jobs are obtained. In this sense, social network analysis points to the fact that individualistic models are insufficient to explain labour market performance because the social context can strongly alter the links between individual capacities and motivations and the expected rewards (Lin, 1999; Portes and Rumbaut, 2010). By so doing, the premises of many network explanations of incorporation into the labour market challenge the liberal view of this process. However, such network explanations usually assume that the professional environment is unstructured and unsegmented (Degenne and Forsé, 1999, p. 114) and tend to ignore the broader institutional framework within which immigrant employment occurs (Waldinger, 1994, p. 4). Very little work has been conducted that takes into account the interplay between on the one hand the structure of opportunities on the demand side, and on the other hand immigrants’ resources (Kloosterman and Rath, 2010).

This article aims to fill this gap by assessing the type of social capital that is most useful for status attainment by immigrants in Spain, given both their human capital and the segmented and ethno-stratified context in which social networks operate. More specifically, it aims to identify what is the most resourceful type of social capital (supportive contacts with the autochthonous population, with co-ethnics living in the country of origin or with co-ethnics living in Spain) for immigrant men and women, who experience (or suffer) the immigrant concentration in the labour market in different ways.

In the next section, the article delves into the literature on immigrant networks and labour market performance. The following methods section presents the network-based methodology adopted here. A personal network survey of 150 Ecuadorian and Moroccan immigrants living in Catalonia, Spain, was carried out using questionnaires on personal networks that retrieved the network of the 30 ‘closest’ contacts for every interviewee. The “Results” section presents the logistic regression analysis that was subsequently performed, showing the relevance of both social capital and the ethno-stratified and gendered structure of the labour market. To conclude, the article reflects on the results obtained in light of the various theoretical debates.

Social networks and labour market attainment

Social networks, defined as social interactions among individuals, are treated as operational and conceptual frameworks that emphasize the relational aspects of social reality and conceptualize social life in terms of the structures of relationships among actors (Carrington and Scott, 2011, p. 6). This perspective takes into account the meso-social dimension of the social structure that provides resources (material, symbolic, informational, etc.) to the agents involved in these networks, thereby both creating and constraining opportunities for action for the individuals embedded in them (Bottero and Crossley, 2011; Bolíbar, 2016). In this sense, following the approach of Bourdieu (1986), Coleman (1988) and Lin (1999), the network of personal relationships may be regarded as the structure where social capital is created, expressed and distributed (Lozares et al., 2011).

Among immigrants, the patterns and structures of informal sociability both reflect and influence the diversity of individual integration processes (Lubbers et al., 2010). Personal networks reveal personal communitiesFootnote 2 (Wellman, 1979; Chua et al., 2011) and express the diversity of individual settlement processes, as they show the different ways in which migrants cut old links with their countries of origin and create new links in their host countries with different kinds of people (Molina et al., 2008; Bolibar et al., 2015). In the empirical study of transnational movements and migratory processes, social networks have been identified as an important resource for migrants’ incorporation into the host society, and consequently as a key element in the process of settlement and social integration (Massey et al., 1993; Portes, 1998; Maya-Jariego, 2001; Cheong et al., 2007; Molina et al., 2008; Lubbers et al., 2010; De Miguel and Tranmer, 2010).

Research focusing in particular on the role of social capital in labour market performance goes back to Granovetter’s seminal research on the strength of weak ties, which pointed out the particular relevance of links to people with no common friends in obtaining new and valuable information on job opportunities. This study opened up a fertile stream of research on the use of contacts in seeking and finding employment, as well as in hiring, recruiting and training. The application of these premises to immigrants’ labour market performance has shown the importance of social networks and social capital in defining the possibilities of migrants being promoted inside a firm (Behtoui and Neergaard, 2010), in determining wage rates (Aguilera and Massey, 2003) and in identifying the sectors in which they are employed (Livingston, 2006). In these regards, there is a debate about what kinds of ties are more useful for immigrants in their endeavours to make progress in the labour market.

Some research points to ties with autochthonous populations as the most valuable for immigrants to succeed in the labour market by attaining a higher employment and occupational status, as by building bridges to the native population such networks span structural holes (Lancee, 2010) that bridge ethnic divides, which are usually also associated with socioeconomic distinctions (Lancee, 2012). In this sense, Lin in particular developed his theory of the strength of position, pointing to the fact that access to and the mobilization of resources embedded in networks, as well as the outcomes of such mobilization in the labour market, depend on the socioeconomic status of personal contacts (1999).

On the other hand, the structure and vitality of ties within immigrant communities that determine immigrants’ ‘reception contexts’ have also been used to explain differences between ethnic groups in their positioning in the structure of labour (Portes and Rumbaut, 2010). Along these lines, Portes and Rumbaut argue that the differential performance between, for example, Mexican and Cuban migrants in the American labour market is not necessarily due to their respective abilities but to the wealth and strength of their community networks, that is, to the resources that are exchanged by means of these networks and the economic organization of ethnic enclaves so as to promote opportunities for mobility through ethnic labour niches in an entirely network-driven way (Portes, 1998, p. 13). Sanders et al. (2002) also identify a negative relation between co-ethnic networks and upward occupational mobility among Asian immigrants in Los Angeles. Other researchers, however, point to the ‘negative effects of social capital’ (Portes, 1998) at work within ethnic-enclaved networks. They might work to concentrate the immigrant workforce within ethnic occupational niches, as the exchange of solidarity and information through social networks might pull migrants into the same situation as their fellow countrymen (Martinez Veiga, 1997). In this sense, the immigrant niche might be a product of the repeated interaction of the networks that link immigrant newcomers to more established compatriots and settle them in those places where their compatriots are already established (Waldinger, 2001; Schrover et al., 2007).

Finally, more recent research rooted in the transnational approach to the migration phenomenon points to the fact that immigrants’ networks are scattered throughout the globe, producing and reproducing transnational social fields in which resources are exchanged in a de-territorialized manner (Levitt and Glick-Schiller, 2004; Vertovec, 2009). Taking this into account, some researchers have also pointed to the value of these transnational ties as a source of business and professional success (Gold, 2001; Kloosterman and Rath, 2010), as they enable the mobilization of resources from different contexts on a global scale.

On top of these three approaches, other developments in the literature have suggested that, in different cultural contexts and in different job categories, factors such as personal reputation or peer-group status may change the kind of ties that different kinds of workers prefer (Degenne and Forsé, 1999). In this regard, the notion of intersectionality can help to explain how social networks often tend to reflect and perpetuate the existing structural inequalities and power relations in a society. Networks may have a different value for different individuals at different times and for different purposes, depending on their social location in the mainstream hierarchical order (Anthias, 2007). In other words, the social valuation of population categories (such as ‘women’ or ‘migrants’) and their articulation or interrelation (e.g., ‘migrant women’), both within a minority ethnic context and within the wider society, can affect the value of their relational resources as well as their ability to use them to create economic capital (Anthias, 2007, 2012). Following these approaches, research has shown that the likelihood of finding a good job as a result of one’s social capital varies along gender and ethnic lines because of the impact of axes of inequality on the capacity to obtain both beneficial information and influential contacts (Trimble and Kmec, 2011, p. 170; Livingston, 2006). Networks are different in form and function for immigrant men and women, as their migration patterns, timing of and reasons for their migration are different, and they are differently embedded in the family sphere (Schover et al., 2007; Bolíbar, 2016). Thus, networks might operate in gendered ways to produce systematic differences in labour market outcomes for men and women immigrants, respectively (Hagan, 1998; Schrover et al., 2007). To sum up, as a result of both unequal access to social capital and the unequal outcomes obtained from it, searching for work through social networks may perpetuate gendered and ethnic labour market inequalities.

Context, data and methods

The data for this study form part of a broader project aimed to assess social cohesion in Catalonia, Spain, through the study of relationships both within and between social groups. The fieldwork was conducted in 2009 and 2010, just after a decade of a significant increase of the foreign-born immigrant population, which in Catalonia rose from 4% in 2000 to 17.5% in 2010 (data from the Municipal Population Register, Statistical Institute of Catalonia). That moment was also particularly marked by the Great Recession, which implied a massive loss of employment and substantial occupational immobility (Arranz et al., 2017). In 2010 the unemployment rate by that time was around 14.8%, but 30% for the immigrant population (data from the Labour Force Survey). The crisis affected the immigrant population particularly hard (Rinken et al., 2018), as they were already in a particularly vulnerable position when it started. Its impact was even worse for those who arrived latest in the Spanish labour market (Rodriguez-Planas and Nollenberger, 2016).

The data were collected by means of a survey of personal networks, administered to a sample of Ecuadorian and Moroccan immigrants (N = 153). In personal networks, the unit of reference is a specific informant (ego) and her relational setting, that is, the individuals with whom she is directly connected (the alters). The survey collected personal networks consisting of 30 contacts, therefore including both strong (close) and weak (distant) ties. To obtain this information, respondents were asked to name 30 people whom they knew by name and with whom they had had contact in the previous 2 years.Footnote 3 Interviews lasted an average of 2 h, as a lot of information about all 30 contacts retrieved by the interviewee was also collected.

The survey followed a non-probability quota sampling method. The sample was stratified in respect of the basic structural characteristics of the Catalan population by city of residence (three cities of different sizes, representing the variations in urbanization in Catalonia), age, gender and origin (country of birth). Multiple methods were used in recruiting participants.Footnote 4

A total of 446 interviews were conducted, 153 of which were with immigrants (77 Moroccans and 76 Ecuadorians) living in Catalonia.Footnote 5 These two ethnic groups were selected because they illustrate well the recent wave of immigration from the Global South that the country has received since the early 2000s. They are the two largest immigrant groups in Spain from outside the EU, and although they both came to Spain for economic reasons, they represent a variety of linguistic, religious and geographical backgrounds.

As shown in Table 1, the sample does not substantially differ from the reference population by sex or age, although the population with the highest educational level is slightly over-represented, especially among the Moroccan population.

Table 1 Comparative descriptive data of the reference population and the sample.
Full size table

After a brief description of immigrants’ networks, logistic regression analysis is used to test the effects of social capital, human capital and sector on the occupational status attained by Moroccans and Ecuadorians in Spain. Further analyses delve into gender differences by considering the interaction effect of each of the three explanatory factors with gender.

Table 2 Summary statistics for study variables.
Full size table
Table 3 Exchange of employment-related support with the contact according to the origin and county of residence of the contact.
Full size table

Acquired occupational status refers to the qualifications of the immigrant’s current or last job. It distinguishes business owners, professionals, qualified workers and workers with management responsibilities from workers without qualifications or only very low ones. The vast majority (98.7%) of those interviewed were wage-earners.

Regarding social capital, three variables were constructed, looking at the size of the sub-set of contacts with whom interviewees had exchanged labour-related supportFootnote 6 according to their ethnic origin and the location of the contact: autochthonous (born in Spain), co-ethnics living in their country of origin (transnational ties), or co-ethnics living in the host country (fellow immigrants in Spain). These variables were dichotomized in the regression analysis based on the median value.

Regarding human capital, the level of education is taken into account (low: basic and compulsory education; medium: secondary post-compulsory and professional education; and high: university degree). Finally, regarding the sector, the analysis takes into account whether the job is located within the ethnic niche: construction, agriculture, hotels and catering, retail and domestic work. In the full model, variables, such as country of birth, length of residence in Spain (in years) and gender are included as controls. Unfortunately, data on language competence or legal status are not available. See Table 2 for summary statistics.

Because of the small sample size, in order to avoid false negatives (LeBel et al., 2008), the significance levels are set at 0.05 (**) and 0.1 (*).

Results

Social networks and job-seeking among Ecuadorian and Moroccan immigrants in Spain

The exchange of employment-related information and assistance through social networks is very important among immigrants. In fact, possibly due to a lack of knowledge of the formal mechanisms of recruitment and because of the undocumented status of some of them, migrants exchange this kind of help with a higher proportion of their contacts (44.4%) than local native-born Spaniards do (34.1%) (p < 0.01).

However, not all contacts in immigrants’ networks provide or exchange this type of social support. In fact, autochthonous ties seem to be the most valuable in these regards, as 54.1% of such ties provide or receive employment-related information or assistance. Conversely, transnational ties with people in the home country appear to be less supportive when it comes to looking for jobs or solving employment-related issues, as only 34.6% of such contacts exchange this type of social support with the interviewees. Finally, employment-related social support is also exchanged with an important proportion (46.4%) of the co-ethnics living in Spain, showing that immigrants from the same place of origin settled in the host country are more able to provide this type of support than those living in the country of origin, although to a lesser extent than autochthonous contacts (Table 3).

The outcome of exchanging employment-related support through social networks

This section explores the role of social capital in the performance of immigrant labour markets. Following the first model, with social capital variables, three sets of models were generated, which gradually add human capital, sector and individual characteristics to the previous models.

The results presented in Table 4 show that social capital is a relevant factor. To be more specific, exchanging labour-related support with autochthonous contacts is associated with being employed in a job with a higher occupational status. These results, moreover, are also found even when controlling for human capital, sector and other individual characteristics, indicating that this is a stable pattern not affected by other possible explanatory factors for immigrants’ labour market performance.

Table 4 Beta coefficients and standard errors (between brackets) of logistic regression analysis testing the effects of social capital, human capital, sector, gender, ethnic group and length of residence in Spain as predictors of immigrants’ occupational status.
Full size table

Regarding human capital, the second model shows that level of education is another good predictor of the occupational status attained. However, this effect ceases to be significant once the sector in which the immigrant works (as presented in model 3) has been controlled for. This suggests that, while those with higher academic qualifications might be in a better position to find a qualified job, if they end up in the immigrant occupational niche, those higher opportunities fade away. Other research also suggests that education may have a lower pay-off for immigrants than for natives in terms of occupational mobility because immigrants with higher levels of education also remain concentrated in sectors with poor job characteristics (Rodríguez-Planas and Nollenberger, 2016), and because their qualifications might not be fully recognized (Miguélez et al., 2013).

Finally, the full model shown in Table 4 indicates that, while ethnic group and length of residence in Spain are not statistically significant, gender is a powerful predictor of immigrants’ occupational status even when the effects of all the other variables are controlled for. In fact, these results confirm the existence of profound differences between the male and female labour forces in their patterns of insertion into the Spanish labour market, which provide fewer opportunities for immigrant women.

In order to determine whether social capital, human capital and ethnic concentration in occupational niches had a different return for immigrant men and women respectively in terms of occupational status, regression analyses setting the interaction effects were performed. These analyses, shown in Table 5, indicate that exchanging employment-related social support with autochthonous contacts remains a significant factor in attaining a higher occupational status, while its interaction with gender is not. This implies that, in line with the results that Kanas et al. (2011) present for Germany, social capital does not give a significant differential return depending on the worker’s gender. It might be noted, however, that segmented analyses show a much stronger effect of having autochthonous contacts in the network among men than among women: while men with this type of social capital work to a much greater extent in qualified occupations than those without it (68.6% vs. 34.1%), for women this asset does not have such a strong impact (only 34.5% work in a qualified occupation, just a few more than the 25.0% who do not have this type of social capital).

Table 5 Beta coefficients and standard errors (between brackets) of logistic regression analysis testing the interactions of social capital, human capital and sector with gender as predictors of immigrants’ occupational status.
Full size table

Working in the immigrant niche, conversely, has a clear differential effect for migrant men and women: the coefficient of working within the immigrant niche is almost zero (and, of course, non-significant), while the interaction between sector and gender appears to be a powerful predictor of the immigrants’ occupational status. This result points to the intersection of ethnic and gender-based axes in the segmentation of the labour market that configures labour-market inequalities: while immigrant men might have greater opportunities for vertical mobility even within the immigrant niche, for immigrant women the immigrant niche becomes a stronger trap, keeping women in undervalued positions within the secondary segment, which seem to have a strong ‘sticky floor’ (Torns, 1999; Parella, 2003).

According to further research, this trend might have been reinforced during the Great Recession, the most remarkable impact of which was the destruction of employment. In the initial stage (2008–2011) this destruction mainly took place in the construction sector, which affected in particular immigrant men. Later on (2011–2014), the exacerbation of the financial crisis and the spending cuts in the public sector also affected immigrant women (Gil-Alonso and Vidal-Coso, 2015). Nevertheless, the Great Recession also produced a worsening effect on employment conditions, hindering immigrants’ occupational mobility (Arranz et al., 2017), particularly for those women who had been regrouped by their male partners, therefore having a shorter period of residence and a migratory project subsidiary to their partners’ career (Moreno-Colom and López-Roldán, 2018). The high unemployment experienced during the crisis also fostered a discriminatory discourse that challenged the legitimacy of immigrants having employment at all, and it also pushed immigrant women to keep or re-occupy the most devalued positions such as intern in domestic service (Parella, 2012).

Conclusion

The data described in this article on the role of social networks in the labour market integration of immigrants clearly refute the assumption of the existence of a meritocratic labour market in which only human capital is rewarded by the market in the form of better jobs. Beyond individual factors, structural factors play an important role in explaining how migrants perform in the labour market, both the well-studied macro-structural elements, such as the occupational structure of the labour market, and the meso-level structure made of social relationships. In this sense, this paper shows that, even when the individual’s human capital and industrial sector are taken into account, the composition of networks of personal relationships creates and constrains employment opportunities for the individuals embedded in them.

Immigrants mobilize their networks in order to obtain information and support for job-seeking to a greater extent than the local Spanish population do. However, personal networks encourage different outcomes for individual immigrants’ depending on the composition of the network. In particular, for Moroccan and Ecuadorian immigrants in Catalonia (Spain) the type of social capital that has been featured as the most valuable for their occupational positioning is that composed of contacts with the native-born Spanish population. In contrast, ties with people living in the country of origin (i.e. transnational networks) are mobilized less in seeking employment-related support, and they are not related at all to the occupation of higher positions in the occupational structure, which questions the theory of the added value of transnational ties. Such ties might be a mechanism of business and professional success among high-status migrants (Gold, 2001) or post-industrial/high-skilled entrepreneurs (Kloosterman and Rath, 2010), but among recent Ecuadorean and Moroccan economic immigrants working mainly in the secondary segment of the labour market, local contacts should be regarded as more useful resources for obtaining labour market rewards than transnational ties.

Finally, the article also suggests that social capital should be understood in light of the wider context and dynamics of labour market segmentation that also play an important role in immigrants’ labour market performance. The article has shown that the specific structure of every labour market needs to be taken into account in order to study the relationship between social capital and its returns. In the Spanish case, the labour market is stratified according to ethnic differences and is crossed by gendered inequalities as well. Therefore, in the context of a sexually segregated labour market, immigrant working women face both ethnic occupational segregation and horizontal and vertical gendered discrimination that places them in undervalued jobs that are closely related to social care. Ethnic and sexual divisions of labour confine immigrant women in a closed and rigid position in the secondary labour market, without possibilities for promotion and with a ‘sticky floor’ (Torns, 1999; Parella, 2003) that hinders their professional development (Grimshaw et al., 2017). In this context, the structure of the labour market seems to make social capital a slightly weaker resource for women immigrants. These results point to the need to consider immigrants’ social capital in light of the intersectionality framework in order to understand how the nature and effects of (ethnic) relationships are crossed by multiple social divisions (Anthias, 2007). Just as human capital might have less salience for immigrant workers as a result of misrecognition and undervaluation of their skills, different forms of social capital also have different effects on social stratification depending on the social location of both the reaching out and reached out actors.

The research presented in this article faces a limitation: it is not possible to assume a one-way causality in the relationship between social capital and labour market performance. On the contrary, there might also be reverse causation (Mouw, 2006), or a complex web of causes and consequences, as the labour sphere is itself a space where networks are developed, allowing immigrants to meet people who might help them in the job search and in their further development. Nevertheless, the article has pointed to associations that provide interesting insights into the dynamics of inequality underlying the position of immigrants in the Spanish labour market. Another limitation has to do with the small N used in the analyses, which does not allow statistical assessment or inference of the results, but only identification of the trends and internal variations in the data. Given the length of questionnaires on personal networks, in personal network surveys small Ns are common, as a result of the trade-off between the number of interviewees and the extension and depth of the information on the network that surrounds every interviewee. In the case of this research, this has provided rich information on the different types of social capital that migrants might have when mobilizing their entire network, including both strong and weak ties.

To conclude, further research may explore this topic in greater depth by combining quantitative data with qualitative data or by working with a longitudinal design. That may allow us to study how networks are created in different life spaces in interaction, and how they influence the ability to obtain resources that may change the opportunity to create new networks. New data are thus needed in order to clarify the causality of the results presented. Further research may also incorporate into the models the distinction between the origin of the human capital of migrants—that is, whether it is acquired in the country of origin or in the host country (Kanas et al., 2011)—and their language skills (Bloch, 2013). Legal status should also be considered in more detail, as rights to residence, work and welfare depending upon immigrants’ status have been shown to contribute to the creation of racialized and gendered socially constructed definitions of work that operate to increase susceptibility to severe exploitation (Lewis et al., 2015).

Nevertheless, the results presented here may help us understand how ethnic and gender inequalities are reinforced in immigrants’ access to and development in the Spanish labour market. It contributes to strengthening an explanation for inequality (Johnson, 2000) that is based on social, cultural, political and economic barriers and structures, rather than on market rewards for responsible individuals’ choices.